Deeptendu Santra
Deeptendu Santra
The unit test is causing problems because it cannot download from UCI archives.
@CarloLucibello should we remove the failing changes? I believe the CI downloading dataset is most likely flagged as bot download.
> I'm not sure why we have an error with the `PTBLM` dataset that we didn't have before, but we can remove it from the CI tests and open a...
The problem is backslashes in lines 36-42 breaking two lines. Breaklines compatibility changed in 1.7 as understood from this [discourse thread](https://discourse.julialang.org/t/breaking-up-long-testset-descriptions/76488/2). It would be better to just use longer lines...
Version Bump for this :0
Done.
Can I work on this?
Been inactive due to Uni. exams, will start working on it today.
Split tensors appear for edge-level tasks in OGB Datasets. The dataset loading for [LinkPropped Datasets](https://ogb.stanford.edu/docs/linkprop/) differs from [GraphPropped](https://ogb.stanford.edu/docs/graphprop/) or [NodePropped](https://ogb.stanford.edu/docs/nodeprop/#ogbn-products). We might need a change of OGB-Dataset APIs. Here are...
Also, API for splits should be consistent for different data sources. eg: Cora and OGBDataset access training masks using different APIs.